ietf
[Top] [All Lists]

Re: Review of draft-ietf-dhc-dhcpv6-prefix-length-hint-issue-05

2017-01-31 10:00:10
Dear Roni,

Thank you for providing a thorough review of this document.

I have a few questions regarding the comments, please see inline.

Best regards,
Tianxiang

2017-01-29 17:02 GMT+08:00 Roni Even 
<roni(_dot_)even(_at_)mail01(_dot_)huawei(_dot_)com>:

Reviewer: Roni Even
Review result: Almost Ready

I am the assigned Gen-ART reviewer for this draft. The General Area
Review Team (Gen-ART) reviews all IETF documents being processed
by the IESG for the IETF Chair.  Please treat these comments just
like any other last call comments.

For more information, please see the FAQ at

<https://trac.ietf.org/trac/gen/wiki/GenArtfaq>.

Document: draft-ietf-dhc-dhcpv6-prefix-length-hint-issue-05
Reviewer: Roni Even
Review Date: 2017-01-29
IETF LC End Date: 2017-02-09
IESG Telechat date: 2017-02-16

Summary: This draft is almost ready for publication as a standard
track RFC.


Major issues:

Minor issues:

1.      I think that this document updates RFC3633 and it should be
mentioned in the title and abstract


[Tianxiang] This document does not change any of the text in RFC 3633, but
specifies the client and server behavior when using the "prefix-length
hint", which was not explained in RFC3633. Would it be suitable to state
"this document updates RFC3633"?


2.      In section 3.1 reference RFC3315 for solicit message and 3.3 for
advertise messge


[Tianxiang] Okay, we'll add reference to RFC3315 in section 3.1 and 3.3
when mentioning Solicit and Advertise messgae.


3.      In section 3.1 solution last paragraph is the order of the two
IAPREFIX options important?


[Tianxiang] There is no strict requirement for the order of the two
options. Would you prefer if we added a sentence to specify this?


4.      In section 3.2 solution why are you suing SHOULD and not MUST in
all recommendations?


[Tianxiang] In this document, we used "MUST" for some of the client
behaviors, as they are mandatory to avoid client failure and necessary if
the client wanted to express its prefix-length requirement.

For server behavior, we used "SHOULD" because this document specifies the
desired server behavior if the server wants to consider the prefix-length
hint, but the server could have its own behavioral policy (e.g. neglect the
prefix-length hint).


5.      In section 3.2 solution, what should the server do if cannot
provide any of the requests


[Tianxiang] If the server could not provide the requested prefix or
prefix-length, it should provide a prefix closest to the prefix-length
hint, as stated in the last sentence of section 3.2.

Are you referring to what the sever should do of it simply could not
provide ANY prefixes to the client? In that case, we could add a sentence
at the end of the section 3.2:

"If the server will not assign any prefixes to any IA_PDs in a subsequent
Request from the client, the server MUST send an Advertise message to the
client as described in section 11.2 of RFC 3633."


6.      In section 3.2 solution last paragraph “SHOULD try to provide” and
in the first paragraph “SHOULD provide” should be the same in both.


[Tianxiang] Thanks for pointing that out, we'll change it to "SHOULD
provide".


7.      In section 3.4 “If the client decided to use  the prefix provided
by the server despite being longer than the  prefix-length hint” yet I
did not see in section 3.2 that the server can provide a longer
prefix.


[Tianxiang] This was mentioned in the last sentence of section 3.2:

"If the requested prefix is not available in the server's prefix pool, and
the client also included a prefix-length hint in the same IA_PD option,
then the server SHOULD try to provide a prefix matching the prefix-length
value, or the prefix with the shortest length possible which is closest to
the prefix-length hint value."


8.      In section 3.5 solution the document use “as close as possible” and
section 3.2 only talk about smaller.


[Tianxiang] This was mentioned in the last sentence of section 3.2.


9.     In section 3.5 solution why offer options 3 and 4. The draft
says that option 3 will break client connections and 4 talks about “a
short non-zero valid-lifetime” which may cause the  client to lose
connection if "too short for the client" since “short” is not an exact
value


[Tianxiang] The idea was to list all the options, and discuss their
consequences. And the server could decide which option to use depending on
its policy.

Option 3 avoids the complexity of handling multiple delegated prefixes,
despite of breaking up all connections. Option 4 allows to server to
configure a valid-lifetime for the old prefix depending on actual
requirements, rather than let the old prefix expire on its own.



Nits/editorial comments:




<Prev in Thread] Current Thread [Next in Thread>